University Athletic Departments Optimize Performance with AI

UF is one of a growing number of schools bringing artificial intelligence into their athletics departments to give coaches and athletes an edge. Just as universities compete to build the biggest stadiums and best training facilities, they are now racing to plant their flags in the next frontier of sports analytics. While it is still early, AI adopters hope the technology will help prevent injuries, optimize performance and uncover winning strategies. 

“AI is having a real impact on sports,” says Ben Alamar, sports analytics consultant and author of Sports Analytics: A Guide for Coaches, Managers and Other Decision Makers. He notes that the uses of AI range from cameras that automatically track and zoom in on wrestlers on the mat to algorithms that analyze every pick-and-roll play in a basketball game. 

“We’re gathering more data than ever on athletes,” Alamar says. “I think AI will dramatically speed up the process of getting value from that data.”

KEEP READING: How one college student is bringing Gen Z into the cybersecurity conversation.

University of Florida’s Athletic Department Bridges Data Silos

Spencer Thomas, director of sports performance and analytics at UF’s University Athletic Association, is focused on using AI to prevent injuries and optimize performance. Currently, he screens athletes for injury potential using a battery of strength and agility assessments, then makes training recommendations based on the results. “In the future, with AI, we can hopefully create more insights,” he says. “Instead of me computing all of this data, we can make the process much faster — and automatic.” 

Researchers use HiPerGator, an NVIDIA-powered supercomputer built on technology donated by NVIDIA co-founder and UF alum Chris Malachowsky, to process their AI workloads. Additionally, the AI initiative relies on sports-specific technology such as inertial measurement unit trackers and plates that measure an athlete’s ground reaction force. The initiative also brings in nutritional data — and even information about academic performance — with the aim of creating a holistic picture of athletes and teams. 

“Athletes are going to Spencer’s lab for performance testing, they’re being monitored during games and practices, and there’s also academic tracking,” Nichols says. “Then there are standard physicals and information about injuries and physical therapy. So, there’s this huge quantity of data.”

With the AI-Powered Athletics initiative in its early stages, Nichols is reluctant to get into specific benefits — both because UF professors have active grant applications related to the initiative and because she doesn’t want to give away the Gators’ game plan to opponents. But she says she is particularly excited about the possibility of uncovering cross-sport insights. 

“We can start asking questions about how training across all of these sports helps maintain health and wellness and lets these players play at peak performance,” Nichols says. “Once you have all of that data, if you can come up with a question, we can answer it. The sky is the limit.”

READ MORE: How legacy application modernization accelerates growth.

AI Gives Vanderbilt University Coaches a New Set of Eyes

Jules White, a computer science professor and senior adviser to the chancellor for generative AI at Vanderbilt University, says athletes and coaches may be able to glean important insights with two simple tools: a camera and ChatGPT. 

Several years ago, White worked with a Vanderbilt Ph.D. student on research using AI to analyze basketball shots. However, he says, that analysis was limited compared with what is possible today. “There’s the AI from before ChatGPT, and then there’s the AI from after,” White says. “It’s a quantum leap, and the vision capabilities of the models are just incredible in their ability to describe scenes and situations.”

White recently demonstrated this difference at a professional baseball game, taking digital photos and asking ChatGPT to analyze the field in real time. The large language model (LLM) identified players by their jersey numbers, read the scoreboard to understand game context, and assessed scoring probabilities based on runner position and situation. 

“The key thing is that it could read the entire field,” White says. 

While the analysis matched “a really good amateur” rather than sophisticated prediction models, White says, the computer vision breakthrough was remarkable. “All you need is a camera on the field, and you know who the players are, what the situation is, everything.” 

The technology has immediate practical applications for teams looking to out-strategize opponents, White says. “You can upload a picture of the field and say, ‘This is who we’re playing this weekend. Here’s how they’re lined up. What do you think is going to happen, and how should we respond?’”

Coaches might even be able to find ways to track specific skills, use AI to visually identify which athletes are struggling and have an LLM send a report to coaches after practice on how to provide those athletes with one-on-one support. “The opportunity is there,” White says. “These things sound like science fiction, but now they are just a matter of engineering.”

RELATED: Here’s how AI-driven analytics help Ohio State manage football stadium crowds.

University of Toledo Requires AI Training for the Athletic Department

Two years ago, University of Toledo Athletics Director Bryan Blair chatted with some of the department’s staffers to learn how they were using AI. “I discovered that there was almost an embarrassment about using AI, because people felt like it was cheating,” Blair recalls. “Other people thought the technology didn’t apply to them, that AI was for younger people or for tech people.” 

To overcome these cultural obstacles, Blair took a simple but powerful step: He mandated AI training for every employee in the department, from the head football coach to the most junior people on the ticket sales staff. 

Rather than pushing any specific goal, Blair wanted to make sure staffers had a baseline understanding of AI and foster a collaborative environment where staff share best practices. Today, employees throughout the department are using tools including Microsoft Copilot and ChatGPT to draft scouting reports more efficiently, help write thank-you notes to donors and generate concept renderings for facilities renovations. 

“If we all adopt these tools and save an hour a week, that’s the equivalent of adding 15 new staff positions in our department,” Blair says. “For a mid-major department like us, that’s a big deal.” 

The department is using online training resources such as LinkedIn Learning and bringing in local AI experts for in-person training sessions, Blair says. As those in the department gain expertise, he hopes more teams will adopt AI for injury prevention and other uses that affect on-field performance. The school’s volleyball coach used AI to prepare for his job interview with the university, and he has since become one of the department’s “super users,” Blair says, using AI to track player performance. 

“Things are evolving quickly,” Blair says. “Those who are at least trying to learn how the technology can be helpful are going to have an advantage.”